ai leaders' safety claims questioned by expert

gary marcus questioned the credibility of sam altman, jensen huang, and bernie sanders on their recent statements about ai safety. marcus argues that altman's claim that the world should trust openai is unconvincing, huang's view that corporations will self-regulate is unrealistic, and sanders' comparison of ai to nuclear war is exaggerated. marcus points out that openai's release of gpt-6 astra, which is less monitorable than its predecessor, suggests a prioritization of market share over safety. public opinion is shifting with nearly two-thirds of americans now seeing ai as a moderate risk to humanity.